In this project, I wanted to look and see if 3-point percentages had increased over the years, as it seems volume certainly has. So I randomly selected 15 players who have played at least 8 seasons, have a 15-minute/game average for their careers and average at least 7.0 points per game. Random.org was used to generate the team to pull from, then again to select a player.

Just a quick look at player shooting per season

Red shows field goal attempts and green shows field goals made

A look into overall shooting finds no significance between the interaction of season, player and field goal attempts

## [1] 0.01729683
## fg_percent ~ season * player * fga

3-Point Shooting for all players over their seasons. We see a shooting volume increase, which help foment the hypothesis that the league might be shooting better from distance.

Red = 3-pointers attempted, green = 3-pointers made, and the blue line indicates 3-point %. For percentages 1 on the graph would = 100% shooting

Now for each player

Red = 3-pointers missed, green = 3-pointers made, and the blue line indicates 3-point %

3-Point percentages for players

0.1 is 10% and 0.5 Is 50% shooting

Analyzing interaction of season and player on 3-point percentages. Significance and residual

## x3p_percent ~ season * player
## [1] 0.07207535

Predictions of the above formala

Predictions in red and 3-point percentage is green

Now, analyzing interaction of season, player, and 3-pointers attempted on 3-point percentages. Significance and residual

## [1] 0.05862911
## 
## Call:
## glm(formula = x3p_percent ~ season * player * x3pa, data = Full_Data)
## 
## Deviance Residuals: 
##       Min         1Q     Median         3Q        Max  
## -0.243148  -0.025831   0.001345   0.024033   0.248441  
## 
## Coefficients:
##                                       Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                           78.48329   20.62749   3.805 0.000218 ***
## season                                -0.03888    0.01023  -3.799 0.000222 ***
## playerBradley Beal                   -96.30801  116.69335  -0.825 0.410708    
## playerChris Paul                    -101.03458   38.05986  -2.655 0.008932 ** 
## playerDamian Lillard                 -67.81600  105.67169  -0.642 0.522158    
## playerDanilo Gallinari               -40.83713   52.87266  -0.772 0.441299    
## playerDeMar DeRozan                  -86.27952   28.36859  -3.041 0.002849 ** 
## playerGordon Hayward                  80.64878   40.92651   1.971 0.050897 .  
## playerIsh Smith                      -68.48368   30.24495  -2.264 0.025212 *  
## playerJeff Teague                   -104.35748   32.94051  -3.168 0.001913 ** 
## playerJonas Valanciunas             -347.11025   61.27221  -5.665 9.02e-08 ***
## playerKemba Walker                  -102.98089   97.96102  -1.051 0.295097    
## playerNikola Vucevic                 -92.46990   41.60846  -2.222 0.027986 *  
## playerRajon Rondo                    -95.96076   29.12070  -3.295 0.001267 ** 
## playerRicky Rubio                    -49.42765  101.79332  -0.486 0.628090    
## playerTobias Harris                  -92.97690   51.25092  -1.814 0.071961 .  
## x3pa                                 381.98977  201.13501   1.899 0.059758 .  
## season:playerBradley Beal              0.04790    0.05791   0.827 0.409661    
## season:playerChris Paul                0.05026    0.01890   2.659 0.008813 ** 
## season:playerDamian Lillard            0.03374    0.05236   0.644 0.520479    
## season:playerDanilo Gallinari          0.02037    0.02628   0.775 0.439719    
## season:playerDeMar DeRozan             0.04285    0.01408   3.044 0.002830 ** 
## season:playerGordon Hayward           -0.04003    0.02032  -1.970 0.050927 .  
## season:playerIsh Smith                 0.03402    0.01502   2.265 0.025140 *  
## season:playerJeff Teague               0.05189    0.01636   3.172 0.001887 ** 
## season:playerJonas Valanciunas         0.17232    0.03041   5.666 9.00e-08 ***
## season:playerKemba Walker              0.05116    0.04868   1.051 0.295238    
## season:playerNikola Vucevic            0.04594    0.02066   2.224 0.027897 *  
## season:playerRajon Rondo               0.04771    0.01447   3.298 0.001258 ** 
## season:playerRicky Rubio               0.02460    0.05053   0.487 0.627210    
## season:playerTobias Harris             0.04622    0.02546   1.816 0.071739 .  
## season:x3pa                           -0.18911    0.09962  -1.898 0.059874 .  
## playerBradley Beal:x3pa             -374.95488  202.10655  -1.855 0.065828 .  
## playerChris Paul:x3pa               -375.57450  201.39750  -1.865 0.064458 .  
## playerDamian Lillard:x3pa           -382.47189  201.75974  -1.896 0.060222 .  
## playerDanilo Gallinari:x3pa         -390.07932  201.33067  -1.938 0.054852 .  
## playerDeMar DeRozan:x3pa            -379.10819  201.71751  -1.879 0.062428 .  
## playerGordon Hayward:x3pa           -421.16052  201.40265  -2.091 0.038462 *  
## playerIsh Smith:x3pa                -376.04543  201.86243  -1.863 0.064736 .  
## playerJeff Teague:x3pa              -374.93642  201.60815  -1.860 0.065184 .  
## playerJonas Valanciunas:x3pa        -129.59884  214.97269  -0.603 0.547651    
## playerKemba Walker:x3pa             -376.76926  201.52225  -1.870 0.063787 .  
## playerNikola Vucevic:x3pa           -374.56593  201.69887  -1.857 0.065566 .  
## playerRajon Rondo:x3pa              -385.34820  201.51045  -1.912 0.058038 .  
## playerRicky Rubio:x3pa              -392.64531  204.83140  -1.917 0.057442 .  
## playerTobias Harris:x3pa            -382.29009  201.45610  -1.898 0.059962 .  
## season:playerBradley Beal:x3pa         0.18562    0.10010   1.854 0.065959 .  
## season:playerChris Paul:x3pa           0.18593    0.09975   1.864 0.064589 .  
## season:playerDamian Lillard:x3pa       0.18935    0.09993   1.895 0.060331 .  
## season:playerDanilo Gallinari:x3pa     0.19313    0.09972   1.937 0.054949 .  
## season:playerDeMar DeRozan:x3pa        0.18770    0.09991   1.879 0.062526 .  
## season:playerGordon Hayward:x3pa       0.20858    0.09975   2.091 0.038483 *  
## season:playerIsh Smith:x3pa            0.18619    0.09998   1.862 0.064826 .  
## season:playerJeff Teague:x3pa          0.18561    0.09986   1.859 0.065318 .  
## season:playerJonas Valanciunas:x3pa    0.06389    0.10649   0.600 0.549554    
## season:playerKemba Walker:x3pa         0.18653    0.09981   1.869 0.063905 .  
## season:playerNikola Vucevic:x3pa       0.18544    0.09990   1.856 0.065677 .  
## season:playerRajon Rondo:x3pa          0.19078    0.09981   1.911 0.058148 .  
## season:playerRicky Rubio:x3pa          0.19440    0.10146   1.916 0.057544 .  
## season:playerTobias Harris:x3pa        0.18927    0.09978   1.897 0.060070 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for gaussian family taken to be 0.005023852)
## 
##     Null deviance: 1.8386  on 189  degrees of freedom
## Residual deviance: 0.6531  on 130  degrees of freedom
##   (4 observations deleted due to missingness)
## AIC: -416.68
## 
## Number of Fisher Scoring iterations: 2
This is the mod that displays great significance of our categories’ correlation to 3-point shooting and the mod shows good residuals compared to other categorical interactions. ‘Players’ shows a lot of significance, as well as certain seasons they played.

3-Point % vs season

Added trend line

3-point percentages and attempts per season

Red = 3-pointers attempted and blue shows the percentages for each season

Predictions using the above interacting formula

Added predictions = red and green shows 3-point % for each season

Predicting into the next 4 future seasons

We see some strong predictions

Interactions of player and minutes per game + season on 3-point %

## x3p_percent ~ player * mp + season
## [1] 0.06984445

This mod shows us significance

Minutes and 3-point % vs Seasons

Given a player shoots good or bad, I’d expect a gain or reduction in minutes.

We see minutes do have an effect 3-point %, but only for certain players. It appears that a consistently decent 3-point % might a good factor into what’s keeping the player playing more minutes. A guy like Andre Drummond, is a traditional big and earns minutes by other methods, but limits spearding the floor like a lot of teams like to do these days.

Interactions of player, position, and season on 3-point %

Another aspect to consider would be that of position. Primary ball-handlers, typicaly the guard, might see different kind of shots then say a center, who will likely only take open threes. But a center’s body type has typically, in the past, not seen them as finesse players with long distant shooting skills
## x3p_percent ~ player * pos * season
## [1] 0.07164746
What could be done in the future is collect mean data from many players at each position and compare those values since there isn’t a big enough pool to draw from within this data-set.